Learning a new language can be challenging, especially if someone wants to do it for professional purposes. Language is one of the biggest barriers for the intermingling of peoples of various communities. Google has undertaken a new one-of-a-kind project that can help lower the language barrier. The project is called Translatotron and it also emphasizes the speaker’s tone, in addition to the spoken words. What’s even better is that this translator would not use any sort of text, and would translate sentences of one language directly into another.
When translating spoken words from one language to another, the words from the first language are converted into text, which is then translated into that of the second language. The text of the second language is then converted into audio. Researchers had been trying for a long time build a translator that carries out direct speech-to-speech translation without the need for the intermediate text. Such translators often use machine learning algorithms. Researchers have also tried to study how the human mind translates languages since understanding human cognition paves the way for sophisticated algorithms.
Although Translatotron is still not perfect, the team at Google that worked on the project is confident of making steady improvements and eventually make direct, real-time translation a reality. With an emphasis on the speaker’s tone, users would not only get a reliable translation but also get to know the inflection of the person in front.